Measuring and Improving Cooling System Performance – Part 3: Diffuser

"Diffuser" is one of those terms that lots of car modifiers use but very few understand. In popular car culture, diffusers are rear bumper garnishes or add-ons that somehow magically create downforce, and you will find lots of these faux diffusers available for purchase at online retailers.

 
You might be surprised to learn that nearly every car manufactured in the last 20 years has at least one actual diffuser on it (most have open rear diffusers now too—"open" because they lack sides and thus are not true diffusers. The ground forms one wall of these diffusers while the opposite wall is a plastic panel under the trunk or, quite often, the muffler). The incorporation of this real diffuser into basic car design has had important effects in improving cooling system capacity and reducing cooling drag in modern cars. It isn't there for downforce, yo; it isn't just for race cars; and it isn't found at the rear of the car. The most important diffuser in your car is at the front, just behind the bumper cover which often forms one wall of the device.

Mass-production diffusers are typically far from smooth, unlike this 2025 Acura Integra lower inlet. This means there is usually room for significant improvement if you don't mind a little fabrication work.

Conservation of Mass (Continuity)
 
So, what are "diffusers" anyway? This is one of those terms you will see people throw around all the time with no clear definition, and usually incorrectly when they do try and identify one. For example, this is not a diffuser:

Wait, this one's a "diffusser." My bad. (Image credit: eBay).

A subsonic diffuser is a type of duct that converts kinetic energy in the flow into static pressure by increasing duct area in the flow direction. (A supersonic diffuser behaves quite differently and has exactly the opposite geometry). Think of it like a "megaphone" shape, with flow entering the small opening and leaving through the large outlet.
 
The reason this works is conservation of mass; that is, if we have a system—here, the diffuser—with flow in and out, we cannot spontaneously create or destroy mass inside the system. Translated from words into math:

This just says that any change in mass inside the system over time must be balanced by mass carried into and out of it. We'll make a simplifying assumption here, and in all subsequent posts, that the cooling system is operating at steady state e.g. at fixed speed and ambient density. In this condition, nothing changes over time—so all time derivatives disappear (the first term in the equation is one such time derivative). We'll also approximate the velocity as one dimensional i.e. we only have velocity in the longitudinal direction in and out. Apply these and solve the remaining surface integral (which tells us how much mass is carried along or convected by the flow) and we get,

or the mass flow through the system—and since mass in must equal mass out, this rate is a constant. If we assume constant density, this "area rule" says that the ratio of inlet to outlet area is the same as the ratio of outlet to inlet velocity,

The flow slows through the diffuser, seen here bending over the rear wheel housing of this Lexus LFA, due to conservation of mass; at equilibrium, mass flow in must equal mass flow out and, assuming constant density, this means a larger area allows less velocity. (Image credit: Reddit).

In order for this to work the duct must not allow mass flow to cross its walls. This is why rear diffusers on cars:

2025 Toyota bZ4X. Even pedestrian cars like this EV sometimes have strakes or vanes on the diffuser panel to improve lateral stability.

...are "open" diffusers: they allow mass flow in and out along the two sides. True diffusers, as on the Lexus above or on any production car with baffles around the heat exchanger designed to prevent airflow around the core, have clearly defined inlets and outlets.

Inlet Diffuser: State 1 to State 2
 
Now back to your car. Total pressure is a measure of the energy available in an air mass; it is the sum of static pressure (internal energy) and dynamic pressure (kinetic energy). If there are no energy losses in the diffuser, the total pressure at the heat exchanger face will equal the total pressure at the grill opening. Of course, this cannot happen in real life since there will always be friction effects in the boundary layer along the diffuser walls, and there may be significant separation in the duct on most cars. Additionally, if your car is older it might not have any ducting at all. In this case, there is nothing preventing air bleeding around the heat exchanger core, reducing mass flow rate through the exchanger.

Our first test will be the measurement of total pressure loss; taking readings between a freestream probe and the heat exchanger core will give us the efficiency of the grill and diffuser combined. Use two total probes (tape a pressure patch to the heat exchanger face, with the port facing forward) and measure the difference.

A total pressure probe is just a tube or port facing forward, into the flow. Separating the total and static probes, as on this SR-71 Blackbird, reduces error in the measurement of dynamic pressure (as we found in a wind tunnel experiment my last semester in school, where total and static ports integrated on the same tube gave more than 10% error compared to separate probes). You can do the same with your total and static probe setup.

Here, I have added a small brass tube, mounted several inches below the static probe, to read total pressure. The difference between these gives dynamic pressure q_∞ I used in Part 2 to normalize p₁, and will use again in just a minute.

Again, you can do this at a range of speeds to see how the total pressure loss varies. Then, divide the measured total pressure loss by freestream dynamic pressure at each speed, giving the efficiency of the diffuser by,

The smaller the difference between the two pressures i.e. the greater the efficiency of the diffuser, the better (we'll see later that smaller drops in total pressure preserve momentum in the flow, and less momentum loss results in less drag from cooling). Here, there is room for improvement since total pressure loss is between 50-60% of freestream dynamic pressure, giving diffuser efficiency just above 40% with slight variation by vehicle speed.
 
Next, let's compare static pressure at the diffuser outlet to the total pressure available there to find the velocity of the air going into the exchanger. In a diffuser with an infinitely large outlet area, the flow will be stopped completely and static pressure will equal total pressure at the diffuser outlet/heat exchanger inlet, a condition called stagnation. If we keep total pressure high by good design of the diffuser (smooth walls and gentle curves) and velocity low, we can minimize internal drag of the cooling system. Measure this by fixing a pressure patch to the heat exchanger face, again pointing forward into the flow, and a tube opening taped to the exchanger face with its end normal to the flow, then recording the difference between them. This difference is the dynamic pressure of the flow q, which is directly related to its velocity u (smaller dynamic pressure means lower velocity and vice versa).

Try to place these probes as close to the center of the heat exchanger stack as possible; on my car, that's about 8" from the bottom and 12" in from each side. This might require loosening or removing the bumper cover to reach.

Use measured q₂ and freestream q_∞ to find the velocity ratio u₂/u_∞ by,

Again, there is some slight variation by speed but not much. This velocity ratio is higher than I would like; we'll see later that internal drag of the cooling system depends on this ratio, and the higher it is, the larger is the internal drag coefficient.

Finally, you should have collected enough data (if you measured atmospheric pressure using a barometer) to calculate the static pressure coefficient at the diffuser outlet/heat exchanger core. The higher this number, the better—since it tells us how much static pressure is recovered from total pressure (the whole point of a diffuser, remember). Calculate this using the same formula as in Part 2:

In the ideal situation, C_P = 1: no total pressure loss (100% efficient diffuser), and stagnation achieved at the heat exchanger core. As I mentioned earlier, this cannot happen in real life but we can try to get as close as possible. C_P < 0.2, as I found on my car, can probably be improved. We'll revisit this in a few posts.

You may notice that I have plotted all test results so far not as absolute values but as ratios or "normalized" coefficients. This often makes information easier to work with and observe patterns in it, as well as building in agnosticism to absolute conditions such as ambient temperature. We'll see later that these ratios—especially the velocity ratio u₂/u_∞—are useful in characterizing cooling system performance.
 
Next time, the heart of the matter: heat exchangers.

Previous: Measuring and Improving Cooling System Performance – Part 2: Inlet